Performance-based contracting based on inputs is particularly challenging in complex settings such as health care, where issues such as joint production, multiple agents in production, and market failures compound the critical contracting concern of multitasking. We analyze the impacts of Rwanda’s national pay-for-performance (P4P) program, where health care facilities were offered performance bonuses based on process measures of health care quality. Using two waves of data from the Rwanda Demographic and Health Surveys, collected before and after the randomized roll-out of P4P, in a difference-in-differences analysis we find that the P4P program improved some rewarded services, had weakly positive impacts on several unrewarded services, and had no impact on health outcomes. We find no evidence of multitasking, and find mixed effects of the program by baseline levels of quality, with most of the improvements seen in the middle quality tier.

This paper evaluates whether health plans in Germany's Social Health Insurance select on an easily observable predictor of risk: geography. To identify plan behavior separately from concurrent demand-side adverse selection, I implement a double-blind audit study in which plans are contacted by fictitious applicants from different locations. I find that plans are less likely to respond and follow-up with applicants from higher-cost regions, such as West Germany. The results suggest that supply-side selection may emerge even in heavily regulated insurance markets. The prospect of risk selection by firms has implications for studies of demand-side selection and regulatory policy in these settings.

We examine how the UK and German health care systems responded to a major cost-saving innovation: the availability of generic simvastatin, a cho¬lesterol-lowering drug. In the German Social Health Insurance, the generic’s entry reduced sales volumes for both branded simvastatin (Zocor) and a close substitute, branded atorvastatin (Lipitor/Sortis). In UK, only the sales of branded simvastatin fell whereas the sales of atorvastatin were mostly unaf¬fected. We trace these experiences to institutional differences in the two health care systems and to the structure of patient cost-sharing in particular.

Improving access to health care and financial protection of the poor is a key concern for policymakers in low- and middle-income countries, but there have been few rigorous program evaluations. The Medical Insurance Program for the Poor in the republic of Georgia provides a free and extensive benefit package and operates through a publicly funded voucher program, enabling beneficiaries to choose their own private insurance company. Eligibility is determined by a proxy means test administered to applicant households. The objective of this study is to evaluate the program's impact on key outcomes including utilization, financial risk protection, and health behavior and management. A dedicated survey of approximately 3500 households around the thresholds was designed to minimize unobserved heterogeneity by sampling clusters with both beneficiary and non-beneficiary households. The research design exploits the sharp discontinuities at two regional eligibility thresholds to estimate local average treatment effects. Results suggest that the program did not affect utilization of health services but decreased mean out-of-pocket expenditures for some groups and reduced the risk of high inpatient expenditures. There are no systematic impacts on health behavior, management of chronic illnesses, and patient satisfaction.

Objective. To evaluate provider responsiveness and beneficiary satisfaction with insurance carriers participating in the Republic of Georgia’s Medical Insurance for the Poor. Study setting. A dedicated survey of approximately 3,500 households in two types of regions – with different eligibility thresholds – in November and December 2008. Study design. Regression-based estimation of responsiveness ratings by beneficiaries and non-beneficiaries of the insurance program and estimation of mean satisfaction scores for beneficiaries. Principal findings. In the high-threshold regions, provider responsiveness toward beneficiaries and non-beneficiaries is comparable. In the low-threshold regions, beneficiary status is associated with lower responsiveness of outpatient providers. Inpatient providers may have become less responsive toward beneficiaries during the program’s transition from public to private administration. While satisfaction of beneficiaries with carriers is above average, there are reports of difficulties obtaining reimbursement and information about benefits. Conclusions. The results suggest that relying on private insurance companies to deliver public programs in middle-income settings may impact provider responsiveness and indicates the need for continuous monitoring and regulation.

This paper examines the effect of systematic self-report bias, the non-random deviation between the self-reported and true values of the same measure. This bias may be constant or variable, and can mislead empirical analyses based on descriptive statistics, program evaluation and instrumental variables estimation. I illustrate these issues with data on self-reported and measured overweight/obesity status, and BMI, height and weight z-scores of public school students in California from 2004 to 2006. I find that the prevalence of overweight/obesity is 2.4–7.6 percentage points lower in self-reported data relative to measured data in the cross-section. A school nutrition policy changed the bias differentially in the treatment and control groups so that program evaluations could find spurious positive or null impacts of the intervention. Potential channels for this effect include improved information and stigma.